3. INTERPRETACIÓN, ANÁLISIS Y DISCUSIÓN
3.1. Características de los titulados: Individuales, académicas y laborales
University
of Cape
Town
University
of Cape
Town
,
CIIA.PTER 4. MUI.TI.,\IODAL TELEMEDICINE INTERCOMMUNICATOR
...
Real Access Crit"ria Criteria Nam" Usage During the Project
~
Local oommunity leader sallc~ioned proj~d
Difficultic:; in adapting ooftware interrace due to Sor',,.cdl.llml ['''.dors general relactanre to critici"", oould he related
to culture. may view criticiflll as sign of disresped
"C"sers familiar with emwl and cellular
Trn"t in te<'hnology tec.hnology 00 had confidence in using new tochnology Socurity beyond soope of projoct
-Projoct may he sewed up if the future !.-[uTI pilot.
Locw oconomic environment is successful [lIJd this conld "pur fnrther developments
.' -.
:\Im:w·,,<:onoUlie cmiromncnt Policy restricted use of VoIP and \ViFi
Di"paritie" in wealth in c.ountry, project region very poor Legi"lation harring VolI' and \ViFt during the
Legal [lIJd regulatory environment project
Project highlighted
lJ "'''
0[' VolP "-nd V>'iFi in a rUl',I! settingDoth VoIP WiFi not supported politically Politicw will and public 8Upport during the project
VoIP now sC1pported since Februar~y 2005 WiFi still restricted
Tallie 5, Thls tahle summanses the maw Rew Access criteria uwl throughout the project
"'
University
of Cape
Town
CHAPTER 4. MULTI-MODAL TELEMEDIGI .... 'E IiVTERGO.\[M[J,'iICATOR
8 Habit~ of Highly Effective ICT Initiativ"s
Habit Name Usage During the Projed
I
Needs assessment rilnrlucted in fir5\ project Do ROme homework, condur-t na" .. 15 c.vrJe as part of first Diagno~i~ phase'Issessmenl
,
Held diocussion groups, interviews wilh andadmini"terro qUC>ltionnaircs to local community members
Allende<.! tclemedicine wmkshop in Umtata \0 learn about telemcdidne in E.a.~tern Cape VisilOO ol.her clinics and hu,pitals in Eastern Capc to oompare situations and w,t a general idea of how heallh care operate" in the provmcc
AllendNl workshop3 wilh CSTR to find out about other devcloplll<"nt initiatives \)<>ing undenaken in South Africa and in Africa UMli soci'i\l.v aware framework to ensure we Implement ~md disseminate best practice foll'J'>md best practice
Identified local champion - the head nm"" of Ensure ownership, local bUY-ln,
nnd a champion
"
..
IsllitwA .. Ahe helper! lacLhtate bUY-ln from oommumty leader who in tum, helped facilitate buy-in
I
Set concrete gml,~md t~ ~mall
achievable "tepl
Critica.lly evalw,te effort~.
report back, adapt as needed
I
frum other community members Milc;toncs set at eac.h Diagno.,is phaseRc-evaluatoo al. each C .. iti~-<ll Re/l&:tion ph~'ge
of socially aWAIe frAme...·ork
RclIection at end of each pro.~\ rycle
I
Reported bar1< to llsers a\ eAch field visit tmd had regula .. conboct by phone and SI>.IS Contlnllcd on uext page
-University
of Cape
Town
CHAPTER 4. MULTI-MODAL TELEMEDICINE INTERCOMMUl,'ICATOR
8 Habits of Highly Effective ICT Initiatives
-Habit Name Usage During the Project
Perronal safety was an issue since the Tsilitwa -Sulenkalllil. regkm has history of .. iolenoo Address key external challenge.; (shootings,hijacking etc)
Dealt with this hy not 8taying overnight in Tsilitwa unless n..coosary
Lobbied DoH to sec benefits of .MuTI They had power \(l make ).luTl 8ustainahle ],Iake it sustainable by incorporating telemedicine "" part
of their provincial strategy for improving healthcare
Project in impoverished area of Involve gronpi5 that are Eastern Cape and involving rural and traditionally e::.:duded isolated commw,ities
.-lahle 6. 1hlS
-
table summarIses the mam proJeclmllnage-mCl,t skills from the Real Acccs.s!Real Impad criteria nsed throughout the project
69
The RAjRl helped ns to widen Our focns 80 that the project was aware of all factors that may have affected the project su~= and we mtempte<:i to address ea<;h crit.erion of RAjRl as best as we muld. Each of the project eycl". will be discussed in more detail in Sections 4.3 and 4.4 in this chapter and in Chapt.er 5. The following section describes the first project cycle in detail. Section 4.4 covers the semnd project cyde.
4 _ 3 Discoyering The Community Needs
This section describes the first cycle of the project with cmphMis On how we discovered the mm-munity needs.
University
of Cape
Town
CHAPTER 4. MULTI-MODAL TELEMEDICTh'E mTEliCOMMUNICATOR
Fil,(ure 11: Tsilitwa dilli"
4.3.1 Diagnosis
On the first cycle of the project, we visited T~ilitwa in AIlb'WIt 2003 to diooover the community needs, Our main ally at thiR point wa,<; the leader of the previous project in Tsilitwa headed by the CSIK He had already built up a trusted re]<l.tionship with Tsilil-wa community and it i" through him that we were introduced to the community memher" who h&d bee:i! mOiSt prominent in the previollil pro,jcct, The rommuuity was initially "uRpicious of OUT intentions but sillL"<l they were familiar with our ally, the CSIIt project leader, they began
w
shuw more trust in UR lIS the firAt trip progre;sed.This oonstituled Our first Diag,wsis stage where we attempted to asooss the cllrrent conditious of the Tsilil-wa project"'" well as identify a problem are" which we could ooutribute to. The guiding questionR uood in the need, assessment can be found in App"ndix A.Ll. We met the head nur-e of the Tsilit"·,, clinic (shown in Figure 11); the twO community member" responsible for maintaining the CSIR \ViFi network aud the current ductor working at the Ne.",i~ Knight hospital in the ueighbouring village of Sulenkama.
The netwurk maintellBnCe RtaJl included" !lChoolteilCher from the school in T"ilitwa and the
University
of Cape
Town
CH"lPTER. 4. MULTI-MOnlL TELEMEDlC1l'.JE LYTERcmL~fUNIGITOH
mana.gl'l" of the Vodacom phone oontainer in Tsilitwa. The Vodacom phone shops are mobile
shipping oom.ainers with about. 5-lO cellular phones for gunerall18e by the publi~ at reasonable call rat"'"' ThL'y arc COmmOn in the Eastern Cape. The net.work maintenance staJl' had be€n contracted and trained by t.he CSlll to manage the WiFi network. The main doctor wru; On leave at this
"t.age bHt we met his comp<1.nion ,,'liO was dae 10 return to Cuba in t.he next mont,h. Both doctors were Cubans on a 3 year contract fOJ the Eastern Cap€ whkh had been negotiated with the South African Doll.
We held a dL<;CU55ioll groap at the hoepital with ail the mmnmnity member" we met, At t,his me€ting, we explained that w-e were underl.aking a rESearch projoct to help build upplication, for the commanity_ Through I,his meeting and ",'yera! informal discass;ons; w€ reali8€d Ilmt, the heaith care staff were experilOlcing difficalties wit,h the telemedicine system which wru; >Lot up by the CSlR Before we elabomte. we fir't deocribe the teiemc'(licine 8ystem placer! in Tsilinva and Sulenkama by the CSlll ..
The CSTR telemedicine ,}",tem Will< d~",igned to ald GOmmHni~"'t,ion for health in the T"ilitwa are.a. The system rdied on a web camem (webclUll) which wa,,.,t ap at the Tsilitwa clinic. Thi' camera containerl a web server and was yiewable from the hospital by snrfulg to its web address, The CSIR implement~'(l a protocol whereby the head nur"" at the clinic coald set up a t.elemedicine appointment wilh I,he doctor at the ho'pitaL Daring thi8 appointment, ,he could com-= with the doct,or u:;ing the VolP phonffi preAt'nt in both locat;oru; ami ml\JlOL'IlVer the Camera to shov; t,he doctor a patient. From this corumltation, it wlls possible for t,he nnr>e In determine if the patient nerooo to be referred to the hospital for treatment or whether the patient coalrl 00 I,reated at the clinic_
Thi, 'Ybtem improved the previoH' prowcol for referrals whereby the narse at the clinic wOHld give patie:nl:; hand written referrals. The patient would then travel to the hospital ami ret11rn to I,he clinic will! a hillld Wl"itt()ll responr<e from I,he noctor. if nece,SlUY. Till, system wa;; not very reliable ,inee often patient' never tmvelled to I,he haspil,al or never returned to the clinic L'Yen if thL')' did tmvel to the hospital. In the new sy.;tem, teleconsaltat,ion bet,ween the nurse and dodor allowed the naffle f\l, the clinic to decide if a patient nc~'(led to 00 referred to the hoepit.a.l. At I,he :;ll.IIIe time, the doctor oould t,hen expect to """ the patient at the h""pital and inform the nar8e
when an examination ",as performed, a'l a folkJ,,"-up.
The telem~'(licine system wa' rle.;igned to obviate the n"'-".! for patients to tmvel to the hospital unneces"-'l1'ily, Since there are 1>0 reliable r.xed phone lines or DECT phon"" in the ure"" the 'y'tem
pro~ided a valuable communication service. TI10 only alternative would he to use handwritten
University
of Cape
Town
CHAPTER 4. MULTHWDAL TELEMEDICINE H';'TERCOMMUb'lCATOR 72
refeual, or mllphones ror communication, Since the DoH doe •• not pay ror cellular calls, this oolution Wa.5 not pr1l.Ctica.i or coot pjfect,ive lor t,he nur>1C8 at Tsilit,,'a clinic or the doctor ttt Kes;ie Knight HO!Ipital in Sulenkama, Also, given the poverty in the surrountling areas, saving a patient the rom, and travel time r.o the ho,pit~l is a siguificant improvement in patient care .. Adtlitionallv, as there was only One medical doctor working at Nessie Knight H",-,pital ttt m(tjt times, patients had to wait in queues for up to s-everal hours berore they "-,,,eivL'd treatmeIlt. Eliminating the wait and trawl time for patiCllt~ that do not require advanN',d treatment, ret!ures the num~ of patients at the hospital and enlHll'es health Care service delivery at the earlie,i point of romact, i.~., the clinic.
Through the tliscussion groups aud inlervie,,'s, we discovered that this
csm
tdemedieine. sY"tem had only been operMional since Kowmbf'r 2(X):1. Also, the system Wail prohlemMic lor several reasons, One of the main problems w8.'l that fwqucnt powL'r failu,,-'S caused network failure and consequently rendered tbe telemedieine, 'y,teIll lmavailable, This meant tbat often t,he wehcatn ~ntlVoIP pbOlK'S were not working and therefore scheduled telemedicinc consultations could not occur.
ALso, in tbis system it was only poosible to have "Y'lc.hronom commnnica.t,ion, i.e., where the raller and rhe, calle" were both available to take p""t in a telepbol>P, call, This was not always pr~ctic.al ""
the nurse and tbe doctol both bad busy 8('hedules, At tbis Stage, it became evident tbat "" could improve the telp~nedicine solution in order r.o make it more robust and bet,rer suited to the local oonditions,
4.3.2 A.ction Planning and Implementation
To ulili~ the know1e,dge, of our allie, to the maximnm, following onr first field trip, in the Adion Plannmg phase of the first proj<JCt cycle, we t\ttended several workshop" and rruJCtings "'ith tbe.
CSIR which included the net,work administrator (and schoolte.ilchn) from T,ilitwa village, who we hy this ,r.age viewed """ ~n ally and the. eurre.nt local champion lor t,he project. Atlditiont\lly.
w<: consulted with Our external monitor. firitlgee.org. Since both the CSIR and Briciges,org were familiar with lCT ror development proje>cts, we felt they would be able to a.5sist us in designing with a oolution for the tdemedicine problem. \\'e a1;o attended a telemedicine, worhhop held at the UniVL'rsity of the ThlllSkci (UNITRA) in Umtata to learn mo:m abunt telemedicine in the EC ama. \Vorking witb these organisations ~nd the commnnity membel"!i or Tsilitwa, ",-e tlecidE'd that ('(lmbining a store and for",Md appro~rb "'ith VoIl' would improVE' r.omrnunict\tion for health in tllC area.
Fraser and McCratb [43[ advocar" an email and store and forward approach t,o telemE'dicine in iwlat-t"d are""'. They ~rgl1e that althongh it m~y not allow real t.ime interaction. it providE'S critical
University
of Cape
Town
CHAPTER 4. MULTI-MODAL l"ELEMEDICINE n-·iTEltCO.UMUNICATOR 73
sp<.ocialist support in the m.ana.gement of difficult casco;. This approach is also economical and the only challenge in implementing these kind of applications is connoctivity. Since connectivity was not an i>6ue for our project, we f<'lt using a Store and f","",·ard approach was apprOj)<iate for the area.
A store and forward approach would allow me88ages to be !lent between the hospital and the clinic while VoIP would allow for phone calls hetween two ;;it"" ~imilar to rhe previous telemedicine
~y~tem. -"'=ges would contain t<'xl, still images and voicemall. :'I-Iessages could J:,;, crcak'<1 and stored at any time and then forwarded if the power and network were available. To deal with the
pool" power supply. we decided the syst<'m would be run on laptops which could provide sevU"aI
hour;, of battery power when a pOWU" failure and subsequent: net:l<lOl"k failure occurs. We developed a VU"y basic paper prototype illustrating these conccpt~ and vi;,ited T;;i]itwa again in December 2003 foc our Action implemen/ation phase of rhe first project cycle.
'Ve uS<.'<1 a show Gild tell approach during a discushlon group with project participants. Our aim wa:; to gather the community's feedback on tbe paper prototype and Our propoo;ed 801ution for w1<'medicine. The network maintenance members " ... re al"" preoont at this di'lCl,ssion group.
During the discussion. we explained Our solution for telemedicine which used VoIP and a store and fOrl',ard approach using the paper prototyp" a~ "vi8ill!.! 3id. 'Ve u.oo the analogy of voic:cmail to deso:;rihe rhe store and forwllId component. Since all of the participanm had cellular phones, they were fumilillJ" with this concept and easily graSp<."<1 the concepts of the proposr.'<1 ,""Imion.
\Ye also explained that rhe sysrem would o .... ercome power probl<'ms that l"<'ndered the previous system unavailable. This was because the nurse could still Create mc&ages during a power failure if rhe system was run on laptop". The"" m<,_"'ages would th<'n be saved and sent when the powe!
resUllli.'<1 and the network was n::storL"<1. OVCl""all, the pap<."I" prototype wa:; very well rc<:eivoo. Both the nill"" and the doctor wU"e positive that the new system would be an impr",""mem on the old sy»l<'m bcealL';c it would aIw allow asyndllonous communicarions.
4.3.3 Ev.~lllation and CrLt.ical TIeftectiQns
In the Action Evaluation ,tw)~ of the Hrst project cycle. we analysed COHill,ent~ and notes from the field trip to confirm whethL"I" "" "'I."I"e propohlng a ,~able solution for r.he network and power problems plaguing the telemedicine system. At this stag<', we found rhat tbe doctor's ma.in problem with the previous telelll€didne system was sch€duling appointments with the nurse. Given his bw;y schedule, it wa:; very difficult for him to guarantee an interruption free hour dedleated to a tdemedicine ~olllmltation. He WdCOmL'<1 the new solut;on feeling positive a.hout its impact on
University
of Cape
Town
CHAPTER 1. MULTI-MODAL TELEMEDJCL'iE L\,THlWOAIAIUNICATOH 71
the patients in the arp-a, The nurse was also plC<l8Cd ,,~th the pro))Otled soluti(}n_ Shp supported the addition of mOlsa~Ol in the new system. In particular, she favoUIed digital pictures (}ver the wpbcam vidro m; she fplt these picture:> were of superior quality, Tn fact, with high r,mution pictures, "he fclt that thpre wm; no nf'ed for vidoo, which due to milllO<'U\wing, poor video quality alld la~, was not practical in many tclcmedicine ooneultations. She was also excited at the prOllpect of using voice messages since nurses at the dinic are often not computer literate and thereforp have poor typing OOlIR.
In the Criti<;al Rcflcchon phas<' of thill firRt "Yd., WP decided that we had ~ucce,"fully id.~ltifif'd
" problem area to work On uoin~ our bottom np approach. Sincp the paper prototype was well If'Ceived, we felt it WM timp to l)l'ocf'ed to implementation of a sof,ware prototype. At the end of the first projPct "Yde, wp al"" idpntifipd thai our main ally and local champion in the commnnity at this stah'e was One of the netwock adminstration stuff, the schoolteacher, Hp was key in helping to arrange and facilitate m""tin~ and \(l help us bnild <l. relationship with other community member", All the details of the first project "Ycle and (}ur reflections wpre dO<'umpnted in a report.
4 .4 The First 1-1uTI prototype
After reRecting (}n the first pwjDCt cycle, we re-evaluated the situation in Teilih.'a ia order to plan our goal" for the oocond cyci<> of the pl'oj!lC1;. We decided that since the paper prototype Wa.5 w<,11 received, it was time to implement a concrete 80ftware prototype to tffit onr ideas_ For this cycle, our goal Wail to dffiign and impi<>ment a telemedicine prototype ba.sed (}n the roq'.lirements we gath-crod from the "",,(}nd field visit. The action plan for the Sf'COnd project eyde oorresponded to the reqnirel1l<'ntR definitio" for aad the design of a software prototype_ Section 4.1,1 detail" the
require---men~e for the prototype and de&Tibes the basic de:lign for this prototype, The implementation of the pro\(ltype is dllicU8:led in Section 4.4.2,
-lA.1 Prototype Requirements and Design
From till> fipld trip" carried out in the first project cycle. :>everal requirements became apparent for a telemedicine software prototype. In general terllli, there was" noed for improved communication between Tsilitwa clinic and thp Nffieie Knight Hospil,al in Sulenkam<l.. A nnmber of factOIR affocted thi" commnnicalion. For inel"aee, there were nO reliable phone lin"" in rhp area al the lime of the project and the electricil,y supply to both the clinic and hoopiral Wail erralic, Addilionally, the previous telemediciue system was being nnderutilif'.eC! d"" to power and nel wock problems
University
of Cape
Town
CHAPTER 4. MULTI-MODAL TELEMEDICINE INTEIlCO.HMUNICATOR
Store and Forward
+
Voice over IP
Figure 12: !>'IuTI required two main components: a store and forward oomponent and a VoIP oomponent. Each of these components in turn have ""veral component" to enable them to carry out the functionality required from them_
rendering it unavailable for large periods. Also. since the old system only offered syncllronous communications, it WI':; difficult to fit in ufiing the "}'slem because of personnel shortages leading to bu"}' "urk schedules.
To improve the telemedicine oolution, a more versatile tool W3.'I required: one which allowed multiple modes of conununication; ooth asynchronous and 8yncllfonous in order to deal with these lmreliable points and allow at Ie""t 3.'Iyuchronous communication if no syuchronous communication
"'as J.>(.AAible. The basic components for the new telemedicine system are shawn in Figure 12: a store and forward component for asynchronous m~ pru;sing and a Voll' component for .ynchronOU8 voice call8.
Watts and Monk [101] a.:naiy""d telemedicine tasks and concluded that yoice and picture!< are essential for most ofth..., tm;\u; . Our prow\ype wm; therefore in mping with general telemedicine trend •. The wftware had to support "}'nchronous voice calls and asynchronous message pa&ling between the dinic and the hospital. These would enable most of the re<juired teLemediciue tasks